QGeoAI logoQGeoAI

Execution and hardware

The Execution and hardware section allows users to configure how inference is executed and to monitor its progress. It covers hardware selection and execution supervision tools.

QGeoAI QPredict device selection

Device selection

The Device block allows users to select the hardware used to run inference.

  • CPU: inference is executed on the processor
  • CUDA: inference is executed on a CUDA-compatible GPU

When CUDA mode is selected, the user can specify the GPU identifier to use. By default, device 0 is selected.

Although inference can run on CPU, using a CUDA-compatible GPU is strongly recommended for optimal performance. A GPU with at least 6 GB of memory is recommended for most use cases.

QGeoAI QPredict inference execution

Running and monitoring inference

Once the configuration is complete, inference is started using the Start prediction button. The Stop prediction button allows the process to be interrupted at any time.

  • Display of a global progress bar
  • Log console showing model information and execution status
  • Loaded model information
  • Inference progress
  • Final execution result: success or failure
  • Potential error messages